Empirical Study of Relationship between Building Materials Indexes and Construction Industry Indexes

博士 === 國立臺灣海洋大學 === 河海工程學系 === 106 === The construction industry is one of the essential pillars of our national economy in Taiwan despite the production value of this industry has been replaced by the high-tech industry. The construction and building products are made from various building material...

Full description

Bibliographic Details
Main Authors: Wang, Yih-Chang, 王益昌
Other Authors: Huang, Ran
Format: Others
Language:en_US
Published: 2018
Online Access:http://ndltd.ncl.edu.tw/handle/6qevt6
Description
Summary:博士 === 國立臺灣海洋大學 === 河海工程學系 === 106 === The construction industry is one of the essential pillars of our national economy in Taiwan despite the production value of this industry has been replaced by the high-tech industry. The construction and building products are made from various building materials such as cement, steel, sand, and gravel. Along with the people's increasing attention for living security and ecological environmental protection, quality and innovation of building materials have been gradually emphasized in recent years. Moreover, the development of construction industry is highly related to real estate expansion. For these reasons, the objective of this dissertation is to examine potential linear and non-linear dynamic paired interactions among indexes of building materials, construction industry and real estate market using Taiwan stock market data. Various time series statistical techniques, including unit root tests, cointegration models, and causality tests are applied to discovery these paired relationships. This empirical work consists of four research topics. The first topic is to study linear and non-linear dynamics between building materials index (cement index and steel index) and construction index. Although linear cointegration tests show no cointegration between these two types of indexes, nonlinear cointegration test based on non-linear threshold error-correction model detects the presence of significant cointegration relationship. Besides, this dissertation finds that non-linear Granger causality between construction index and cement index is bi-directional but steel index exerts non-linear Granger causality on construction index in both the long run and the short run. The second topic is to focus on the dynamics between building materials indexes and real estate market index (real estate investment trust (REIT) index is used as a proxy variable). Using four different cointegration techniques, it fails to find evidence of cointegration. The standard linear Granger tests detect a bi-directional causality between cement and REIT index returns, and a unidirectional non-linear causality from REIT to steel index returns. However, the non-linear Granger causality tests only detect the existence of a unidirectional non-linear causality from REIT to cement or steel index returns. The third topic is to examine the dynamics between construction index and real estate market index. This dissertation still does not find evidence of the linear or non-linear cointegration relationship between them. However, the result of linear Granger tests and non-linear Granger tests show existence of a bi-directional causality between these two index returns. The final topic is to explore the dynamics between real estate market index and aggregate stock market index in the Taiwan stock markets. The results of linear and non-linear cointegration tests show the non-existence of cointegration relationship. The standard Granger tests show weak evidence of Granger causality from the REIT to stock index returns but fail to provide evidence of reversed Granger causality. Moreover, the results from non-linear Granger tests indicate the non-existence of non-linear causality between real estate market and stock market. The findings of this dissertation have important practical implications for investors and policymakers seeking to promote investment strategy and sustainable development in construction industry chain.